Icon Legend

This session is not in your schedule.

This session is in your schedule. Click again to remove it.

Biomedical Imaging and Instrumentation

Biomedical Imaging and Instrumentation - Poster Session B

Poster F5 - Determining The Viability of Point-of-Care Ultrasound for Quantifying Vascular Access Site Stenosis

Thursday, October 24, 2024

2:45 PM - 3:45 PM EST

Location: Exhibit Hall E, F & G

Presenting Author(s)

Grant K. Becker

Summer Intern - APT Center
Case Western Reserve University
Pleasantville, New York, United States

Co-Author(s)

Angel Ramirez

Undergraduate Researcher
Case Western Reserve University, United States

Primary Investigator(s)

Steve Majerus, PhD (he/him/his)

Assistant Professor
Case Western Reserve University, United States

Introduction: Vascular access is crucial for hemodialysis patients with arteriovenous fistulas (AVFs) and arteriovenous grafts (AVGs), the two most common access types. AVFs and AVGs have high blood flow rates (>600 ml/min) but are susceptible to stenosis and thrombosis, leading to failure. Vascular access dysfunction leads to increased rates of morbidity, mortality, and cost of care for dialysis patients. Surveillance of all hemodialysis patients for thrombosis risk is currently performed with angiography or Doppler duplex ultrasound. Because these tools generally require skilled interpretation and are performed outside the dialysis center, they are too costly for widespread surveillance of asymptomatic (but potentially at-risk) vascular access sites.
Point of Care Ultrasound (POCUS) may be used to assess vascular access thrombosis risk as a bedside alternative to other surveillance methods. While ultrasound is used for vascular access surveillance, POCUS leverages highly streamlined protocols and quantitative approaches instead of radiologists' comprehensive workup. POCUS has been adopted in internal medicine (for example, cardiac echocardiography), but its utility for assessing vascular access is not fully established. In this study, we assessed how accurately POCUS could quantify the degree of stenosis (DOS) under controlled in vitro flow models. We hypothesized that POCUS-derived measurements of hemodynamics, coupled with machine learning data analysis, may provide accurate quantification of the DOS. This study aimed to look at the practicality and variation of ultrasound measurements under laboratory conditions in preparation for future trials.

Materials and

Methods: A GE LOGIQ E9 Ultrasound machine was used to record six measurements on stenosis phantoms using B-mode and pulsed wave Doppler. The measurements taken were Time-Averaged Mean Flow Rate (Vm-cm/s), Flow Volume (FV-ml/min), End-Diastolic Velocity (EDV-cm/s), Resistance Index (RI), Pulsatility Index (PI), and Peak-Systolic Velocity (PSV-cm/s). Eight vascular phantoms were cast from Smooth-On EcoFlex 00-10 two-part silicone to mimic 10%-80% DOS in increments of 10%. Each phantom contained two 60 mm ID parallel silicone tubes running down the center with a suture to create two stenotic lesions per phantom. An O-Proline suture with the knot tied on the bottom of the tube created the lesion and reduced shadowing during imaging. A pulsatile flow setup was created using a diaphragm pump (SHURFLO 4008-101-A65), silicone and latex tubing, and pressure and flow sensors. The pump was controlled using LabView and powered from an adjustable power supply (B&K Precision 9111) to induce flows with adjustable pulsatility and flow rate. Calibrated blood mimicking Doppler fluid (BMF) (CIRS Model 769DF) was pumped through the phantoms. Measurements were collected at eleven points spaced 1 cm apart along each phantom vessel at three flow levels (6V, 9V, 12V). At each point, the six measurements (Vm, FV, EDV, RI, PI, PSV) were recorded seven times to characterize the inherent variability.

Results, Conclusions, and Discussions:

Results: The time-averaged mean flow rate (Vm) was plotted against all DOS at location 2 (figure 1). A 5th-order polynomial fit yielded a correlation coefficient of 0.865 with a positive correlation between Vm and DOS. At locations 2 and 4, the Pulsatility Index (PI) data was plotted across all DOS (Figure 2). A 5th-order polynomial fit yielded a correlation coefficient of 0.995 and 0.985 for locations 2 and 4, respectively. A negative correlation between PI and DOS was found.

Discussion: This study tested POCUS under highly controlled in vitro settings, mainly to assess how reproducible and accurate certain ultrasound-derived measures of hemodynamics are. The measured data suggested that PSV, EDV, RI, and FV exhibited significant variability and may be less useful in a machine-learning approach to multi-parameter estimation of DOS.

Some of the investigated measurements may be useful in determining the DOS, and therefore thrombosis risk, in AVFs or AVGs. The Pulsatility Index (PI) data showed the most correlation with DOS and was relatively independent of recording location after the stenotic region. The PI data at locations 2 and 4 revealed an inverse correlation between PI and DOS. An increase in Vm also correlated with an increase in DOS at location 2. This finding was consistent with previous studies and guidelines that indicate DOS over 50% is hemodynamically significant, inducing the so-called Tardus Parvus waveform. This is characterized by a drawn-out waveform with a large time-averaged mean flow rate.

Conclusion: One key finding in this study was that no single measurement was sufficiently descriptive of DOS over variable flow rates, even with highly controlled vascular geometries and focal stenotic lesions. This suggests multiple variables may be needed to determine the DOS accurately, especially given the expected clinical variability. Future work will include applying machine learning algorithms to fuse these multiple measurements together into a single predictor of DOS. The study's future work hopes to find a combination of measurements that can be used with POCUS to make for more efficient and accurate quantifications of stenosis levels for hemodialysis patients.

Acknowledgements (Optional): Thank you to the Louis Stokes APT Center, Dr. Allison Hess-Dunning, and Dr. Jeffrey R. Capadona for organizing the Wen H. Ko summer internship program, which allowed me to learn so much and grow as a student and scientist. Thank you to Athena Veizi for your help with the project. Thank you to Dr. Steve Majerus for being such a great mentor and for his help with the project.